Objective To assess the association of anthropometric measurements with neonatal complications in macrosomic newborns of non-diabetic mothers.

Design Retrospective cohort study.

Patients All liveborn, singleton, full term newborns with birth weight ≥4000 g born to non-diabetic mothers at a tertiary medical centre in 1995–2005 (n=2766, study group) were matched to the next born, healthy, full term infant with a birth weight of 3000–4000 g (control group). Exclusion criteria were multiple birth, congenital infection, major malformations and pregnancy complications.

Intervention Data collection by file review.

Outcome measures Complication rates were compared between study and control groups and between symmetric and asymmetric macrosomic newborns, defined by weight/length ratio (WLR), Body Mass Index and Ponderal Index.

Results The 2766 non-diabetic macrosomic infants identified were matched to 2766 control infants. The macrosomic group had higher rates of hypoglycaemia (1.2% vs 0.5%, p=0.008), transient tachypnoea of the newborn (1.5% vs 0.5%, p<0.001), hyperthermia (0.6% vs 0.1%, p=0.012), and birth trauma (2% vs 0.7%, p<0.001), with no cases of symptomatic polycythaemia, and only one case of hypoglycaemia. Hypoglycaemia was positively associated with birth weight. It was significantly higher in the asymmetric than the symmetric macrosomic newborns, defined by WLR (1.7% vs 0.3%, p<0.001).

Conclusions Macrosomic infants of non-diabetic mothers are at increased risk of neonatal complications. However, routine measurements of haematocrit and calcium may not be necessary. Symmetric macrosomic infants (by WLR) have a similar rate of hypoglycaemia as normal-weight infants. Thus, repeat glucose measurements in symmetric macrosomic infants are not justified.